Stanford experts say COVID-19 death estimates may be too high

Daily Post Correspondent

Although some have estimated that COVID-19 could kill millions of people, those projections — which are shaping government responses such as shelter-at-home orders — might be too high by orders of magnitude, according to two Stanford researchers.

The World Health Organization and others have estimated that 2% to 4% of people with confirmed cases of COVID-19 will die, figures that could be used to project that 2 million to 4 million people in the U.S. would die if 100 million Americans get the respiratory disease, Dr. Eran Bendavid and Dr. Jay Bhattacharya said in an opinion column in the Wall Street Journal on March 24.

But Bendavid and Bhattacharya called those figures “deeply flawed.” The way to find the true death rate would be to look at fatalities as a percentage of people who have been infected with the new coronavirus — not just those who are tested and become confirmed cases, they said.

And if the number of people infected with the virus is much larger than the number of confirmed cases, projections of fatalities would drop substantially, Bendavid and Bhattacharya said. That could mean COVID-19 would kill 20,000 people rather than 2 million, they added.

“If it’s true that the novel coronavirus would kill millions without shelter-in-place orders and quarantines, then the extraordinary measures being carried out in cities and states around the country are surely justified,” they wrote. “But there’s little evidence to confirm that premise — and projections of the death toll could plausibly be orders of magnitude too high.”

Bendavid is an associate professor of medicine at Stanford and a core faculty member in the Center for Health Policy. Bhattacharya is a professor of medicine and director of the Program on Medical Outcomes.

In addition to running in the WSJ, which has a paywall, excerpts of their column are posted on the website of Stanford’s Freeman Spogli Institute for International Studies, which doesn’t charge.

In an email responding to questions from the Daily Post, Bhattacharya said he isn’t claiming that the widespread shelter-at-home orders are unnecessary.

But more information is needed to determine what types of protective measures are needed, he said. That includes how many people have been infected with the virus, how widely it is spreading, and how many people are becoming seriously ill or dying from it.

Bendavid and Bhattacharya are organizing a study in California to collect some of that information. They are planning to use blood tests to see how many people have been infected with the new coronavirus.

The blood test looks for antibodies that the body has made in response to an infection with the virus. The test determines whether someone has been infected, regardless of how ill they’ve become from it or if they’ve since recovered. This type of research is known as a seroprevalence study or a serosurvey.

In contrast, tests that involve nasal swabs from a patient look for the virus’ genetic material, a sign of current infection.

Bhattacharya said the antibody tests are just now becoming available in the U.S. He said the researchers hope to start testing in Santa Clara and Los Angeles counties this week.

“The facts to date are consistent with a wide range of uncertainty regarding the fatality rate from COVID-19,” Bendavid said in an email. “We desperately need a population-representative estimate of the seroprevalence of the disease so we can reduce that uncertainty and make better policy on the basis of our improved knowledge.”

Bendavid’s message was an auto-reply that said “we are struggling to keep up with inquiries.” The researchers have raised some money for their study but are looking for additional funding, he said.

In a report this month, the World Health Organization said the COVID-19 death rate is between 3% and 4% of reported cases. WHO acknowledged that the death rate as a percentage of infections, rather than reported cases, would be lower.

WHO last month called for prioritizing COVID-19 research including seroprevalence studies, such as the one the Stanford researchers are planning. The studies have reportedly begun in other countries, including China.

Bendavid and Bhattacharya cited a few examples where data is available regarding the coronavirus infection rate. When people were evacuated from Wuhan, China, in late January, 2,433 of the evacuees were tested on arrival, quarantined, and tested again 14 days later. Of those, 0.9% tested positive for the coronavirus, the researchers said.

Applying the 0.9% positive rate to the population of the Wuhan area produces an estimate of 20 million people infected with the virus — and a death rate that’s at least 10-fold lower than figures based on number of reported cases, they said.

“If we’re right about the limited scale of the epidemic, then measures focused on older populations and hospitals are sensible,” Bendavid and Bhattacharya wrote.

Other researchers are analyzing the expected spread of COVID-19. At the Institute for Health Metrics and Evaluation at the University of Washington School of Medicine, researchers are projecting deaths from the respiratory disease to hit a peak of 2,341 per day in the U.S. on April 14. They project that total COVID-19 deaths in the U.S. will plateau over the summer, reaching 81,114 by Aug. 4.

The projections “assume the continuation of strong social distancing measures and other protective measures,” they said.

In San Mateo County, the Health Department is tracking COVID-19 data, but it’s still unclear what the trajectory of the disease will be, County Manager Mike Callagy said in a conference call with local elected officials on Friday. San Carlos City Councilman Mark Olbert provided a recap of the conference call on his website.

The county is discussing extending its emergency regulations beyond the current April 7 end date, Olbert said.


  1. If the response to COVID-19 has been exaggerated, what recourse does an ordinary citizen have? The county health officer isn’t elected. The governor isn’t up for re-election for two years, when this crisis will be forgotten. The government got away with a big one here.

    • Don’t call it the “govt”.

      The democratic states have 20k of the 29k deaths from covid-19.

      Heart attack deaths and stroke deaths have “disappeared” for some reason and are now covid-19 deaths.

      Even with the deliberate accounting errors going on we just now broke the seasonal flu death total which peaked a MONTH ago.

      Check out the growing list of 90+ year olds that have survived covid-19.How “deadly” is it again?

      Democrats need to be demolished in the upcoming elections that is the recourse.
      Send the lying bums packing…

  2. The hospitals in Santa Clara County were supposed to be hit with a surge of COVID-19 patients starting today. Didn’t happen. Hospitalizations were the same as last week. Very few COVID-19 in El Camino’s ICU department.

  3. “Applying the 0.9% positive rate to the population of the Wuhan area produces an estimate of 20 million people infected with the virus — and a death rate that’s at least 10-fold lower than figures based on number of reported cases, they said.”

    Why not apply it to the population of Wuhan instead? That gives ~100k, which is reasonably near the number of detected cases.

  4. The CDC says we’ll have between 24,000 and 63,000 flu deaths this flu season (which starts in September). So far we’ve had only 12,000 COVID-19 deaths. I don’t understand why the news media isn’t reporting the flu death toll, and focusing on how the nation is grappling with this killer. And why did we shutdown our economy and our society for something that isn’t even half as deadly as the flu?

    • Flu Deaths = 40K in 1 YEAR with no social distancing
      COVID-19 Deaths = 14K IN 1 MONTH with extreme social distancing to reduce number of deaths

      Very clear that COVID-19 is much more dangerous than Flu.

  5. This part is incorrect – “Applying the 0.9% positive rate to the population of the Wuhan area produces an estimate of 20 million people infected with the virus”

    It’s not 20m infections, but 0.9% of the 20m = 180k. Still far higher that reported cases of 5-10k.

    “After 14 days, the percentage who tested positive was 0.9%. If this was the prevalence in the greater Wuhan area on Jan. 31, then, with a population of about 20 million, greater Wuhan had 178,000 infections, about 30-fold more than the number of reported cases”

    • We don’t really know the actual true # of cases in China. I’ve heard there is suspicion the numbers were under reported.

      Also we’re all guessing at the true death rate, without unlimited testing. the 2-4% is of patients with symptoms tested, which will still hold if we still continue to only test people with symptoms and ignore anyone asymptomatic…..

      if one assumes 10 million are infected, some SUBSET will be tested with the current testing scheme which requires symptoms, and the death rate would still be 2-4% of those. That of course would be less than 200-400k….

      gotta compare apples to apples and not oranges….

      • Agreed there appears under reporting in Wuhan. My comment was meant to correct a error in this piece which garbled the numbers from the original WSJ article by Bendavid. They did not estimate 20m infected but 178,000, back in late Jan. I included the direct quote from WSJ for comparison.

  6. People need to wake up! Use your brains and read what’s being discovered. It doesn’t take higher education to analyze the numbers and facts to see this was grossly miscalculated. Isolation/quarantine has gone on for too long. The new problem is that governors in some states now love the control and will “kill” their people and economy and will not bat an eye while doing so. The media is responsible for fear mongering. It’s too bad they can’t be accountable for their LIES!!

Comments are closed.